English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Prior work has shown that combining bootstrap imputation with tree-basedmachine learning variable selection methods can provide good performancesachievable on fully observed data when covariate and outcome data are missingat random (MAR). This approach however is computationally expensive, especiallyon large-scale datasets. We propose an inference-based method, called RR-BART,which leverages the likelihood-based Bayesian machine learning technique,Bayesian additive regression trees, and uses Rubin's rule to combine theestimates and variances of the variable importance measures on multiply imputeddatasets for variable selection in the presence of MAR data. We conduct arepresentative simulation study to investigate the practical operatingcharacteristics of RR-BART, and compare it with the bootstrap imputation basedmethods. We further demonstrate the methods via a case study of risk factorsfor 3-year incidence of metabolic syndrome among middle-aged women using datafrom the Study of Women's Health Across the Nation (SWAN). The simulation studysuggests that even in complex conditions of nonlinearity and nonadditivity witha large percentage of missingness, RR-BART can reasonably recover bothprediction and variable selection performances, achievable on the fullyobserved data. RR-BART provides the best performance that the bootstrapimputation based methods can achieve with the optimal selection thresholdvalue. In addition, RR-BART demonstrates a substantially stronger ability ofdetecting discrete predictors. Furthermore, RR-BART offers substantialcomputational savings. When implemented on the SWAN data, RR-BART adds to theliterature by selecting a set of predictors that had been less commonlyidentified as risk factors but had substantial biological justifications.

A flexible approach for variable selection in large-scale healthcare database studies with missing covariate and outcome data